The broad objective of the proposed research is to evaluate and understand the impact of a national tobacco control policy. Specifically, the proposed research will evaluate the psychosocial and behavioral effects of the new graphic cigarette package warning labels that were introduced in Canada in December 2000. These labels contain vivid photographs of the negative health consequences of cigarette smoking, with text messages inside the package that include efficacy messages for reducing the threat depicted on the package. In a quasi-experimental design, the participants will be a total of 12,000 students at 9 high schools in Canada and 6 high schools in the United States. Schools in the United States are matched so that they are similar to the Canadian schools. Baseline data confirm that the Canadian and U.S. schools do not differ in smoking behavior. Students will participate in a 30-minute classroom administered survey 2 times a year for 3 academic years. The first wave of data collection was conducted prior to the introduction of the new warning labels. There are six specific aims: (1) to examine whether the graphic warning labels have any effects on smoking behavior; (2) to examine whether the graphic warning labels enhance label-relevant psychosocial variables (e.g., label salience); (3) to examine whether the graphic warning labels diminish general psychosocial variables that are related to smoking behavior; (4) to examine whether the effects of the graphic warning labels will vary as a function of exposure to the labels; (5) to examine whether the effects of the graphic warning labels will vary as a function of individual difference variables such as sensation seeking and time perspective; and (6) to examine whether the effects of the graphic warning labels on smoking behavior are mediated by variables that have been identified by theory to be important in predicting and understanding smoking behavior. Data-analytic methods will be employed that take into account the clustered nature of the data (e.g., random-effects regression models). The proposed research takes advantage of a "natural experiment" in an important tobacco control policy, and thus has the potential to gain insight into the effectiveness of these and other efforts by governments to reduce tobacco use, and could provide an empirical basis for future tobacco control policies in the U.S. and other countries. Moreover, the simultaneous focus on policy evaluation and theory testing may yield insights into the underlying mechanisms that would explain how and why the policy achieved its desired effects (or why it failed to achieve its desired effects).